Endometriosis is an enigmatic disease that affects roughly ۵–۱۰% of women of reproductive age women and is a complex syndrome consisting of multiple vague symptoms such as pelvic pain and infertility, yet, its underlying molecular mechanism is still largely unknown.To unveil the underlying molecular mechanisms of endometriosis (including micro-RNAs and their potential target genes and TFs), we used a systems biology approach, and we constructed two gene-regulatory networks (GRN) for up and down-regulated genes (DEG) and one
PPI network.Methodology We firstly compiled endometriosis-associated genes in women affected with Endometriosis. A comprehensive list of differentially expressed genes (DEGs) was extracted from seven different microarray studies and from three public disease-gene databases. Seven endometriosis-related published data were extracted from GEO and Array Express databases, then normalized and analyzed with geWorkbench software. The genes which were available in both types of lists (expression and text-mined) were considered as endometriosis candidate genes.Three public databases including CTD (http://ctdbase.org/),Cormine ( https://www.coremine.com/), and Malacards (https://www.malacards.org/)were also used to extract text-mined Endometriosis-related genes. Besides miRNAs as post-transcriptional regulators of gene expression, transcription factors (TFs) are also the main regulators at the transcriptional level. To construct the two GRNs, we explored the regulatory relationships between miRNAs, target genes, and TFs. Next, two gene-regulatory networks consisting of Endometriosis-related genes and their known human TFs and microRNAs were constructed for up and down-regulated genes separately. We first incorporated together four types of regulation to construct a transcription factor-microRNA-target gene network. These extracted regulatory relationships (TF͢ miRNA, miRNA ͢ Gene, TF ͢ Gene, and miRNA ͢ TF) which were extracted from five different databases (including mirTarBase, miRecord, TransFact Jasper, TRRUST, and TransmiR databases) can be combined in ۱۳ different types of ۳-node motifs. Then, ۳-node regulatory motif types were detected in the resulting network using FANMOD software. Finally, each motif type was evaluated for its significance using random network generation. The random networks were built ۱۰۰۰ times to compare with the original input network. When randomizing the network in a constant local model, edges with the same relationships were exchanged five times. Z-scores were also computed for all the motif types. They indicated the frequency of motifs observed in the real network minus the mean of their occurrence in the random network divided by the standard deviation. Finally, from multiple types of motifs in the dumping results, ۳-node motif types having Z-score >۲.۰ and p-value < ۰.۰۵ were considered significant.